EXTENDING THE NATIONAL CHILDCARE SCHEME TO CHILDMINDERS: COST AND DISTRIBUTIONAL EFFECT

A series of studies have found that, prior to the introduction of the National Childcare Scheme (NCS), parents in Ireland faced some of the highest childcare costs among OECD countries. To improve childcare affordability, in 2019 the Government introduced the NCS, which awards universal and means-tested childcare subsidies to families using registered childcare. In this report we estimate the potential cost and distributional impact of extending the NCS to a form of mostly unregistered childcare: childminders who care for children in the childminder’s home . We estimate that this reform would increase the cost of the NCS to the Exchequer by € 35 – 121 million per annum (8 per cent to 29 per cent of the current cost), depending on take-up of the subsidy. Such a reform would decrease the cost of childminder-based care by € 100 per month, per child, and would mainly benefit middle-income households. This could also have wider impacts, on both the demand for centre-based childcare and mothers’ labour supply.

Growing Up in Ireland (GUI) is funded by the Department of Children, Equality, Disability, Integration and Youth (DCEDIY).It is managed by DCEDIY in association with the Central Statistics Office (CSO).Results in this report are based on analyses of data from Research Microdata Files provided by the Central Statistics Office (CSO).Neither the CSO nor DCEDIY take any responsibility for the views expressed or the outputs generated from these analyses.This paper has been accepted for publication by the Institute, which does not itself take institutional policy positions.The paper has been peer reviewed prior to publication.The authors are solely responsible for the content and the views expressed.

SECTION 1 Introduction
A series of studies have found that parents in Ireland face some of the highest childcare costs in the OECD (OECD, 2007; OECD, 2015; OECD, 2020; OECD, 2021).According to Motiejunaite-Schulmeister et al. (2019), the average monthly fee for full-time care for children under three in Ireland was €771 in 2019, among the highest in the EU.The relative burden of childcare costs varies by household type.Russell et al. (2018) showed that lower-income families and lone parents spend relatively more of their disposable income on childcare.
In 2019, the Government introduced the National Childcare Scheme (NCS) to ensure equal opportunity childcare, improve childcare affordability and lower the barriers to labour market participation. 1 The NCS provides universal and meanstested childcare subsidies to families using registered childcare.It replaced all previously available childcare subsidies, except the Early Childhood Care and Education (ECCE) scheme.The subsidy is awarded at a rate of €1.40-5.10 per hour of registered childcare used up to a maximum of 45 hours per week.The hourly rate is determined by the age of the child, household income and parental work and education status. 2 Registered childcare, in practice, equates to centre-based (formal) care and excludes childminders and nannies.Childminders are care-givers who mind children in their own home and are typically self-employed.Nannies care for children in the child's home and are usually employed by the parents.Both childminders and nannies are termed 'informal' care-givers as they are not registered with Tusla. 3 Childminder care is, however, the second most widely-used form of paid, non-parental childcare in Ireland (Russell et al., 2016; Byrne and  O'Toole, 2016).In theory, childminders can register with Tusla and become eligible for NCS subsidies.In practice, very few do so, which leaves the informal childcare market largely unregulated. 4evious research has examined the effect of the NCS on childcare affordability and mothers' labour supply.Doorley et al. (2021) estimate that in 2017, prior to the introduction of the NCS, out-of-pocket childcare costs represented an average of 10 per cent of families' disposable income, although a substantial minority of households faced childcare costs of more than 20 per cent of their disposable income.Simulating the effect of the NCS on childcare affordability, the authors estimated that it was equivalent to a cash transfer of 8 per cent of disposable income for the lowest-income households, falling to 1-3 per cent for middle-and high-income households.Doorley et al. (2023) showed that the introduction and recent expansion of the NCS is likely to put pressure on formal childcare facilities by increasing the demand for subsidised, formal childcare over informal, non-subsidised childcare.Doorley et al. (2023) estimate that expanding the NCS to informal care would decrease the demand for formal childcare by around 20 percentage points.
Through the National Action Plan for Childminding (DCEDIY, 2021), the Government has committed to extending regulation and Tusla registration to all non-relative childminders on a phased basis.Legislation and childminder-specific regulations are currently being advanced by the Department of Children, Equality, Disability, Integration and Youth (DCEDIY) that will allow childminders to register with Tusla and take part in the NCS.The National Action Plan commits to a transition period, during which childminders will be encouraged to register but will not be required to do so.
This report evaluates the potential extension of the NCS to childminders in terms of the Exchequer cost and the distributional impact.We use There is a well-documented pattern of imperfect take-up of cash and non-cash benefits both nationally and internationally (Bhargava and Manoli, 2015; Doorley  and Kakoulidou, 2023).The DCEDIY will allow for a three-year transition period, where childminders will be encouraged to register with Tusla, but will not be required to do so.These two factors are likely to result in a low take-up of the NCS for childminders, particularly at the beginning of the three-year period.For this reason, we present several scenarios in which only a proportion of children in childminder care become eligible for the NCS.

SECTION 2 Disaggregating informal childcare types
We use the Irish Survey of Income and Living Conditions (SILC) linked to SWITCH and the Growing Up in Ireland (GUI) Survey.The GUI survey collects detailed information on the type of childcare used for the study child.In the latest infant wave, study children are all aged nine years.SILC, on the other hand collects information on childcare for all children in the household, but with less detail on the type of childcare. 6In particular, SILC groups nannies -who mind children in the child's home -and childminders -who mind children in their own home -together under the heading 'informal childcare'.The SILC data are not granular enough to distinguish between relative and non-relative nanny or childminder care, so a limitation of this analysis is that we are likely to overestimate the number of childminders eligible for the National Childcare Scheme (NCS) under the proposed policy reform.
The first step of our analysis is to estimate the relative split between children cared for by childminders and those cared for by nannies, as the NCS will be extended to the former but not the latter (DCEDIY, 2021).
We use Wave 5 of the GUI '08 cohort, which was collected in 2017, when study children were nine years old (McNamara et al., 2021).Childcare questions concern regular care arrangements made for the study child during term time.We define formal childcare as preschool or afterschool care in a group setting, organised homework club, afterschool facility, special needs facility or activity camps (Appendix A).Childminder care includes paid relatives/friends or childminders in their own home. 7In practice, to be subject to regulation and Tusla registration, the childminder will need to care for at least one non-relative child.
Nanny care includes paid relatives/friends or childminders in the child's home or au pair/nanny.Among children in paid childcare, 31 per cent of children in the sample use formal pre-or after-school care, while 69 per cent use informal care (Table 2).Out of the 69 per cent of children who use informal childcare, the majority use a childminder (39 per cent).
We use the GUI data to model the probability that a child in informal childcare uses a nanny versus a childminder.The dependent variable is equal to one if the child is 6 The SWITCH model reweights the SILC data to be representative of the population using, among other controls, population projections from the Census.The age bands used in the reweighting procedure are gender specific and, while the SWITCH model simulates the number of children quite accurately (Keane et al, 2023), it may not be accurate on children of specific ages.Future model development will examine the feasibility of reweighting by more granular age bands.7 Our estimates from the 2017 GUI survey indicates that 26 per cent of families using childminders are paying relatives to undertake this role.We do not know, however, if these relatives are also caring for any non-relative children.Care by unpaid relatives is recorded separately and not included in the 'childminder' category.
Extending the National Childcare Scheme to childminders | 5 cared for by a nanny and zero if they use a childminder.Explanatory variables,   , are limited to those that are comparable between the SILC and GUI datasets.
Table 1 shows the results of a number of specifications for equation 1, estimated using a linear probability model, with standard errors clustered on the household level.The choice of explanatory variables is partly motivated by variables that are common between the GUI and SILC datasets and partly motivated by observable differences between the sample of nanny and childminder users in the GUI data (see Table A3  The probability of using a nanny rather than a childminder is positively associated with the weekly cost and hours of care.The age of the youngest child is negatively associated with the probability of using a nanny while the age of the eldest child is positively associated. 9Household income and mothers' work hours are positively associated with the probability of using a nanny but these variables are missing for some observations and are not statistically significant in specifications ( 2) and (3) so our preferred specification is specification (1).

8
The weekly cost of childcare is inflated to 2019 levels using CPI in order to be comparable to reported childcare cost in SILC.9 We tried a number of different specifications for this model, including adding the number of children, a lone parent dummy and commuting time, but most additional explanatory variables that we added (i) were not statistically significant or had tiny coefficients or (ii) did not materially affect the split between childminders/nannies.We opted to use the simplest specification (in column 1) for transparency.
We use the estimated coefficients from column (1) in Table 1 to predict whether children in 'informal' care in the SILC dataset are likely to be cared for by a childminder or a nanny.Due to the different age profile of children in the GUI survey compared to the SILC survey, we see a different baseline split between formal childcare and informal childcare in the two datasets (Table 2).The use of formal childcare is more frequent for children surveyed by SILC than by GUI.This is likely due to the younger average age of children in formal compared to informal care.For example, the majority of three and four years olds receive some formal childcare due to the provision by the ECCE scheme of free preschool.As children of this age are in the SILC dataset but not the GUI dataset, we expect to see different proportions of children in each dataset in formal and informal care. 10formal care (69 per cent) -predominantly childminders (39 per cent) -is the most frequent type of care for the nine-year-old study children observed in the GUI data, while formal childcare (69 per cent) is the most frequent type of care for children observed in the SILC data.Employing the estimated model coefficients from column (1) in Table 1, and using the SILC dataset, we predict the proportion of informal childcare provided by nannies versus childminders.The key assumption underlying this imputation is that, even if nine year olds have systematically different childcare provision to younger or older children, in the case of informal childcare, the split between nannies and childminders is related to the same observables -cost, hours of care and age of children in the household -for children of all ages.We estimate that, among the 31 per cent of children in informal care in the SILC dataset, fewer than one-fifth are cared for by nannies while the rest are cared for by childminders (column SILC 1).
As a robustness check, we also use data from the Ipsos CSP survey, which was conducted by DCEDIY in 2022 and which collected data on a representative sample of nearly 1,000 children under 15 years of age.Using the relative proportions of nannies and childminders within informal care types reported in that survey for children of school age (5-14 years) and not of school age (0-4 years), we randomly assign the same proportional split to children observed in the SILC data.Our results (shown in column SILC 2) are very similar to predictions using the GUI data (column SILC 1).Within informal care provision, we estimate that around 10 per cent of children are cared for by a nanny, with the remaining 90 per cent cared for by a childminder.
In what follows, we use the SILC 1 estimates, obtained by mapping the nanny/childminder split observed in the GUI data via the model coefficients shown in column 1 of Table 1.Note: Authors' calculations using children in 2017 GUI data, children 0-12 years in 2019 SILC and children 0-14 years in 2022 Ipsos Childcare Survey of Parents.SILC 1 imputes the split between nannies and childminders for informal care types using the coefficients from equation 1, estimated using the GUI data.SILC 2 imputes the split between nannies and childminders for informal care types using the relative proportions from the IPSOS MRBI data for children 0-4 years and children 5-14 years, randomly attributed to children 0-4 years and 5-12 years in informal care in the SILC data.

Descriptive statistics
Having estimated the proportion of children in informal care in the Survey of Income and Living Conditions (SILC) data cared for by childminders versus nannies, we next present the characteristics of individual children under 13 years of age in each type of childcare settings in the 2019 SILC data (Table 3).
Most children under 13 years of age (79 per cent) are not in childcare.These children live in households with relatively low equivalised household income and low parental labour supply, indicating that they are probably cared for by a parent.A further 5 per cent of children are in unpaid care, by a relative or friend.The labour supply of their parents is higher than that of children who are not in childcare but the average hours of unpaid care is low, at 12 per month.
Around 10 per cent of children are in a formal childcare setting such as a crèche, Montessori or afterschool club.A small proportion of these use only 15 hours of care per week in the context of the free preschool scheme, the Early Childhood Care and Education (ECCE) scheme.The average equivalised disposable income of centre-based (but not ECCE-only) children is relatively high and they use 99 hours of care per month on average.The average cost of their care is €454 per month, for which they receive an average subsidy of €169.
Around 5 per cent of children are cared for by a childminder only.These children tend to have an older age profile and to have parents who work longer hours.They receive 58 hours of care per month, on average, at a cost of €318 per month.
Just 1 per cent of children are cared for by a nanny only.These children tend to have more siblings and parents who work longer hours.The average hours of care is similar to the case of childminders, at 101 per month but the cost is much higher, at €1,063 per month.
Finally, 1 per cent of children are cared by a combination of formal and informal care (nanny or childminder).These children live in households with the highest average equivalised disposable income and high parental labour supply.The average hours of care used by these children is the highest of all of the groups, at 144 per month at a cost of €665 per month, on average.These children receive an average National Childcare Scheme (NCS) subsidy (for their hours in a formal setting) of €85 per month.
Children cared for by a childminder only, or by a combination of centre-based and childminder care, are newly eligible for the NCS under the proposed legislative Extending the National Childcare Scheme to childminders | 9 change.Both groups of children live in households with relatively high average income and high parental labour supply.In the next section we examine the effects of NCS extension for these two groups.Note: Authors' calculations using SWITCH linked to 2019 SILC data.Sample is children <13 in 2019 SILC data.Income is equivalised using the national scale.The split between nanny and childminder in SILC data is predicted using coefficients from equation 1, estimated using the '08 cohort at age 9 GUI data.Greyed cells show means based on a relatively small sample size of 30-50 observations and should be interpreted with caution.
Extending the National Childcare Scheme to childminders | 11 SECTION 4

Extending the NCS to childminders
We simulate several scenarios in which the National Childcare Scheme (NCS) is extended to children cared for by childminders.Proposed legislation will allow for a three-year transition period, where childminders are allowed to register with Tusla, but are not legally required to do so.Take-up may also be incomplete if parents do not systematically claim the subsidy.11Therefore, at the beginning of the three-year period, we expect relatively low take-up, which will subsequently rise over time. 12We model three scenarios for registration of childminders and related take-up of the NCS.In the first scenario, 25 per cent of households using childminders avail of the NCS for these hours of care.This corresponds to a hypothetical scenario in which few childminders register with Tusla and/or few households apply for the NCS for their eligible hours of childminder care.A further two scenarios increase this proportion to 50 per cent and 75 per cent, simulating an increase in childminder registration/take-up of the NCS.In the final scenario, all households using childminders are assumed to avail of the NCS for their eligible hours of care.
In the counterfactuals with imperfect take-up of the NCS, we use five random draws per scenario to assign households to take up or not take up the subsidy.We then average the policy effects over the five random draws for each take-up scenario.In each case, the simulated scenario is compared to a baseline situation where the current (2023) parameters of the NCS are in place and the NCS is fully taken up by families using formal childcare who are entitled to it.Note: Authors' calculations using SWITCH linked to 2019 SILC.The share of children cared for by childminders is estimated using coefficients from equation 1 estimated using GUI data (column (1) in Table 1).For each take-up scenario (25%, 50%, 75% of childminders), results are averaged over five random draws for that scenario.The total cost in the baseline scenario (no childminders eligible for NCS) is based on the 2023 parameters of the NCS applied to all eligible children in formal childcare.
Table 4 shows the annual cost of extending the NCS to childminders and the average monthly subsidy for children cared for by childminders.The annual cost of the NCS increases by between 8 per cent and 29 per cent, depending on the number of childminders who take up the scheme.This corresponds to an annual extra spend of €35-122 million per annum.
The average monthly subsidy for children cared for by childminders is €96 in the full take-up scenario.This is less than the current average subsidy, of €169, for children in centre-based care (Table 4).The lower average hours of care for children with a childminder, combined with their higher average equivalised household disposable income, means that these children on average benefit from a lower subsidy.
Table 5 shows the number of extra recipients of the NCS and the increased Exchequer expenditure, split between the universal and means-tested components of the NCS, assuming 100 per cent take-up of the subsidy by children cared for by childminders. 13The number of universal subsidy recipients increases by 54,000 while the number of recipients of the means-tested subsidy (either standard or universal) increases by 26,000 (around half of the first figure).The corresponding increases in expenditure are 37 per cent for the universal subsidy and 23 per cent for the means-tested subsidy.Extending the NCS to childminders, therefore, primarily increases expenditure on, and receipt of, the universal subsidy.This is in line with the observation that families of children cared for by childminders have relatively high average disposable income, limiting their eligibility for means-tested subsidies.Figure 1 shows the distributional effect of extending the NCS to childminders in the four take-up scenarios.Households are ranked from the lowest income fifth (quintile) to the highest income fifth, based on their equivalised disposable income.All households, and not just those with children, are represented in this chart to show the targeting of the policy reform across the income distribution.For this reason, the average gains as a proportion of disposable income appear very low as they are diluted by the zero-gain experienced by all households without children cared for by a childminder.

FIGURE 1 DISTRIBUTIONAL IMPACT OF EXTENDING THE NCS TO CHILDMINDERS
Note: Authors' calculations using SWITCH linked to 2019 SILC.Quintiles are based on equivalised disposable household income using the national equivalence scale.The four scenarios correspond to simulated take-up by childminders of the NCS (25%, 50%, 75% and 100%).appendix), which is strongly progressive.Currently, households in quintile 2 benefit most from the NCS, while household in quintiles 4 and 5 benefit least.Figure 1 shows that extending the NCS to childminders would result in additional gains which are larger for households in quintile 3, followed by households in quintiles 4, 5 and 2. There are no gains in the bottom quintile of the income distribution as there are no households with children cared for by a childminder in this income quintile. 14 Behavioural responses to this reform might include switching behaviour between formal and informal childcare.This would change the pattern of gains observed here.

Conclusions
The NCS is an important policy instrument to reduce the cost of childcare for families in Ireland and increase female labour supply.In this report, we evaluate the proposed extension of the National Childcare Scheme (NCS) to childminders.We find that this would increase the total cost of the NCS to the Exchequer by 8 per cent to 29 per cent, depending on scheme take up.During the initial phase of the transition period, costs are likely to be at the lower end of this estimate.We find that the subsidy would be worth nearly €100 per month on average for children cared for by childminders according to current usage.This represents close to 30 per cent of the current cost of care for these children.
Previous research has shown that the NCS is progressive (Doorley et al., 2021).The subsidy represents a higher proportion of disposable income for low-income households, compared to high-income households.A reform to the NCS which targets childminders would benefit middle-income households the most, reflecting the fact that households that use childminders have higher average disposable income than households using formal or unpaid care.
The extension of the NCS to forms of currently unregistered care, specifically childminders, could have wider positive societal impacts by increasing the demand for these types of care, alleviating the pressure on centre-based childcare.Previous research has found that children cared by childminders (or parents) have a small advantage over those in centre-based care in terms of health, socio-emotional difficulties and vocabulary (McGinnity et al., 2013, 2015; Russell et al., 2016).Recent research by Doorley et al. (2023) shows that if the price of informal care fell, more parents would choose this type of care, reflecting an innate preference of many parents for non-centre-based care.Furthermore, as informal childcare remains largely unregulated, with only a few childminders registered with Tusla, incentivising childminders to register could have significant effects on quality of care and health and safety.By reducing the cost of an alternative form of childcare, the policy could also increase the labour supply of mothers, a group that demonstrates relatively high responsiveness to financial incentives to work (Doorley et al., 2023).
Some limitations to our analysis include our inability to separate childminders who are relatives from childminders who are not.For this reason, we are likely to overestimate the cost of extending the NCS to childminders.Additionally, we have not accounted for potential behavioural effects.Increased demand for childminders as a result of the subsidy could lead to supply shortages or price increases.

APPENDIX
In SWITCH, childcare costs and hours are based on self-reported responses in 2019 SILC, the data underlying the SWITCH model.The SILC data contains information on hours used per week for each type of childcare in a 'usual week'.The average weekly cost of each type of childcare is also collected.

Note:
Authors' calculations using SWITCH linked to 2019 SILC.Quintiles are based on equivalised disposable household income using the national equivalence scale.

Figure 1
Figure1shows that most of the gain from extending the NCS to childminders goes to the middle of the income distribution.This pattern of gains contrasts with the distributional effect of the current version of the NCS (shown in FigureA1in the Paid relative (or family friend) in his/her own home Table of contents | i
in the appendix): weekly childcare cost per child, weekly childcare hours, age of the youngest and eldest child in the household, annual equivalised household income and mothers' hours of work. 8

The
Growing Up in Ireland (GUI) survey collects information on childcare type, hours and cost for the study child.For the 2008 cohort at nine years of age (collected in 2017), the information collected is as follows:What is the main type of out-of-school care, if any, that you currently use during term time for child?15This question includes all care types organised/provided by a public or private body.16For this variable there are direct arrangements between the carer and the parents: 'Professional' childminder shall be understood as a person for whom looking after the child represents a job of work or paid activity.The term 'professional' does not necessarily indicate a qualification or concern quality of care.Babysitters and 'au pair' are also included here.Extending the National Childcare Scheme to childminders | 19

TABLE A2 CHILDCARE VARIABLES IN GUI '08 COHORT AT 9 YEARS TABLE A3 DESCRIPTIVE STATISTICS, GUI INFANT COHORT, WAVE 5 Formal childcare Informal childcare Nanny only Minders only Total
Sample is children in the GUI '08 cohort age 9 who are in centre-based care or cared for by either a nanny or a childminder, with siblings under age 14.PCG = Primary care giver.HH = Household.
Informal -childminder Paid childminder in your own homeInformal -nanny Paid childminder in his/her own home